In the name of God

Humanoid Robot
S.Vahid Hashemi , Mahdi Moradinasab
Mahdi Najmabadi , Mohsen Zangeneh
Ali Abbasi , Saeed Abdoshshah
Semnan University
www.iranfanavaran.netfirms.com
iranfanavaran@gmail.com

.Abstract
This paper describes the designing of a humanoid robot called "Dara" in Semnan
University by team IRAN FAN AVARAN which wants to participate in the Humanoid
League of Robocup 2006 in Germany. The main purpose of this project is to build a
humanoid robot to simulate the behavior of human and going to use this generation of
robots in some special works instead of human.
This mechatronic project contains the implementation of Control Algorithms, Mechanical
Design, and Electronic Control Boards under operating multiple processors together.
In designing of this robot has tried to use some mechanism, control algorithms,
mechanical structures and electrical exciters which didn't has still used in similar projects.

1. INTRODUCTION
As this project has defined, it contains a
biped robot with 140cm height that can
see environment, hear voices and
process them. This robot can recognize
and execute some pantomime or vocal
orders, like pointing to right, pointing to
left, saying "forward", "sit down",
"stop", "Dara"
The team started research and theoretical
working on this project in 2002. After
several months' collection of information
about same projects in the world and
designing the mechanical and electronic
system, we started to construct the
primary robot in 2003.Picture number
one shows this robot. Because of
appearance some problems in
mechanical construction step and
causing later difficulties we decided to
design a new structure. In the following

we describe specification of the new

humanoid robot, called "Dara".

Picture 1

2. System in a shot
The Electronic system is a set of
processors that work with each other
coordinately.
According to the picture 2, one
embedded computer is responsible for
processing the environments
information which is detecting with
camera and microphone. The computer
distinguishes the task, and then

Picture 2

Sends related commands according to

our protocol to Main Processor.
Main processor is a DSP from
TMS320c54 family that receives the
commands from computer. Also it
receives the feedbacks from joints and
sensors from by an AVR
microcontroller. After processing the
data, DSP send the necessary comment
to some slave AVR micro controllers
which are responsible for movement all
parts of the robot.

3. Mechanical Structure
The height of this robot is almost 140cm.
consisting of twenty three joints or
Degrees Of Freedom: six DOFs per leg
on position-wise three DOFs at each hip,
one DOF at each knee, and two DOFs at
each ankle; one at trunk, four at each
hand and two at neck , picture 3,shows
total 23 degrees of freedom, without
fingers freedom.
For designing structure and modeling of
each part of the robot we used CATIA
software. Picture 4 shows some parts of
the robot in Catia environment.

and Force sensors for some parts of the

robot used in balance calculation and
some reactions.

Picture 4

5. Hardware in Detail

wrist

The communication between processors

in our system has its own protocols. The
data communication between computer
and DSP is Parallel port combining with
handshaking!
The data communication between AVR
and DSP is selected synchronous serial.
Because of difference between
synchronous serial protocol of DSP and
universal synchronous serial protocol of
AVR an Altera FPGA, 7000 series is
used to convert two protocols each other.
Pictures 5 & 6 show the processor
boards, DSP and local AVRs which has
designed.

Picture 3

4. Sensors
One of the important parts for industrial
system is sensor for optimizing control
system and make system more
intelligent.
Except camera and microphone as
vision&voice receivers, we used Touch
Picture 5

DSP, the main processor is responsible

calculating the Center Of Mass (COM),
ZMP and management & synchronize
overall the three control parts.

7. Actuators

Picture 6
For recognition degrees, speeds and
acceleration of each joint we used linear
potentiometer. The analogue signals sent
to AVR microcontrollers, every part of
the robot has its local AVR. After
converting the signals to digital, the data
store in a special database to use in
control algorithms.
The sampling period of joints is 1.5ms.

6. Control
The control section consist three parts:
Balance control, velocity control and
position control.
Essential for the walking of a biped
robot is the balance control. An efficient
algorithm of balance control is necessary
to achieve dynamic walking control.
Walking motion of biped can be
determined by the hip trajectory and the
swing foot trajectory. The stability can
be characterized by the Zero Moment
Point (ZMP) criterion.
In this work a fuzzy algorithm is
proposed to control the lateral plane
movements for the robot. The algorithm
is based on a fuzzy PD control
algorithm, were the output gain is
incremental and sensible to error. This
feature guaranties a smooth balance
control response and a correct
interaction between the walking
algorithm and the lateral plane control.

Actuators which are used in this robot

are three types:
7-1. DC geared motor with 35v and
20A max and torque of 12 Nm, for the
legs and hip
7-2. DC geared motor with 12v and
6A max and torque of 8.5 Nm, for hands
and trunk
7-3. Stepping motor with 24v, 0.3A
and torque of 5 kg-cm, for neck

8. Simulation
Modeling and simulation of the robot
can greatly assist in making progress on
all levels of design, implementation and
operation.
In this project some parts of simulation
and modeling the robot dynamics as well
as measures for stability and
performance of legs locomotion has
done by Matlab GUI.

9. Image processing
One of the most important part of the
robot is its eyes, because of distinguish
its position and environment and also for
navigation.
We take a picture from a CCD camera
and process our image in the industrial
computer.
Because of need to distance
recognition we should have two
eyes, this allows us to see two
different images from two points of
view. Then by analyzing the
differences between these two
images we are Able to deduct some
characteristics of the 3D world

around us. In the picture 7 the

structure of two cameras is shown.

10. Voice recognition

We have made specific software for
processing voices and recognition some
vocal order like "stop", "sit down",
"right"," left"
This goal has achieved by Delphi
software. Because of need to considering
so many factors for correct recognition
we defined special users in the software
to can talk and order the robot. So the
robot only executes vocal orders of few
people.

Picture 7
Two main algorithms which we have
used are Chain code and SUSAN. When
dealing with a region or object, several
compact representations is available that
can facilitate manipulation and
measurements on the object. In each
case we assume that we begin with an
image representation of the object as
shown in picture 8. Several techniques
exist to represent the region or object by
describing its contour.

Picture8
We implemented image processing
algorithms by Delphi software

11. Conclusion
As described, based on our four years
scientific researches, experiments and
experiences, we have designed Dara, the
humanoid robot.
It has 23 degrees of freedom and has
new design in mechanism & electronic
system as compared to other humanoid
robots, because many parts of this
project have designed by our own team.
Also it has designed with low cost
components and actuators.
The robot is equipped with vision and
hearing, it is fuzzy controlled and can
improve and develop for using in human
world.
Now we are making it ready for
Robocup 2006 competition.